dawngriffiths - gbv.de · geometrie, binomial and poisson distributions. popcorn machine xvi...
TRANSCRIPT
Head First Statistics
Wouldn't it be dreamy ifthere was a statistics book thatwas more fun than an overdue trip
to the dentist? But it's probablyjust a fantasy ...
Dawn Griffiths
Q'REILLY®Beijing • Cambridge • Köln • Sebastopol • Taipei • Tokyo
table of contents
Table of Contents (Summary)Int ro XXVII
1 Visua lizing In for m ation : First Impressions
2 M easuring C en trat Tenden cy: The l'vfiddle Wrlj' 45
3 Measuri ng Sp read: Poicer Ranges 83
4 C alculat ing Pro ba bilities: Taking Chanres 127
5 Discrete Proba bility Distri butions: Ma nage l'jllr Expectatious 197
6 Permuta tions and Combi nation s: Making Artangements 24 1
7 G eometrie, Binom ial, a nd Po isson Di stribution s: Kee/Jillg Things Disa ete 269
8 Nor m al Dist ribution: BeiugNormal 325
9 No r mal Di stribut ion Part I I: Beyond Normal 36 1
10 l Jsing Starisrical Sampling: 7äking Sampies 41 5
11 Estimating Your Popula tion : /vlakillg Predutions 44 j
12 Co nsrrucring Confidcnce lnrervals: Gucssing tritl: COlifide/lrf 487
13 U sing Hyporhesis Te sts: Look at the Emdeure 52 1
14 The Chi Sq uare Distribution: There's SOIl/e1hillg Goiiu; Oll 567
15 Correlarion a nd Re gression : Whal's kl)' Lilie? 605
Append ix i: Top Tell Thillgs We ou« Cover 643
11 Ap pendix ii: Statistus Tables 65 7
Table of Contents (the real thing)Intro
Your brain on statistics. Here yau are trying to learn something, while
here your brain is doing you a favor by making sure the learning doesn't stick. Your
brain 's thinking, "Sett er leave room for more important things, like which wild
animal s to avoid and whether naked snowboarding is a bad idea" So how da you
trick your brain into thinking that your life depends on knowing statistics?
Who is thi s book for?
We know wh at you 'r e th inking
M etacognition
Bend your brain int o subrnission
Read me
The technical review team
Acknowledgm ents
XXVII I
XXI X
XX X I
XXXlll
XXX IV
XXXVI
XXX VII
ix
table of contents
1v'lsual'lz'lnh' Int9rmettl9n
First Impressions
Can't tell your facts from your figures?Statisties help you make sense of confus inq sets of data . They make the
eomplex simple. And when you've found out what's really going on, you
need a way of visualizing it and telling everyone else. So if you want to
piek the best ehart for the job , grab your eoat, paek your best slide rule, and
join us on a ride to Statsville .
Staristics <I re eve rywhere
But why learn staristics?
A tale of two charts
2
3
4-
10
10
11
12
13
14
18
19
39
3
34
20
26
'2 7
28
35
Company Profit per Month
Deali ng with grouped dat a
M ak e a histogram
Step I: Find the bar width s
Step 2: Find the bar heigh ts
Srep 3: D rall' your chart
Inr rodu cing cumularive lrequcu cy
Dr awing the cumulative Irequency gr aph
Choosin g rhe right charr
The hum ble p ie cha rt
Bar cha rts ca u a llow for more accura cy
Verrical ba r charts
H orizont al bar cha rts
It's a matt er of scale
U sin g Irequency scale s
DeaJing with multiple sers of data
Catcgories vs, numbers
~ 2 S.!!"0
2.',...0.. 2.3c
E 2.2
!. 2.1jEe 2.0"- Jul Aug Sep Oe' Nov Oe<
Month
See whot r meon , t heprof it 's obout t hesame ea ch month.
No, th isprof it's omoz ing.Look ot it soor!
Ju l Aug Sep Oc t Nov Dec
Montn.--~-_
Company Profit per Month
1
~ 2S~"0
2.0,...0.. 1.5c
~ 1.0
!. 0.5jE0 0 .0.t
x
2
table of contents
meQsurlng centraI tendencj
The Middle Way
Sometimes you just need to get to the heart of the matter.It can be difficult to see patterns and trends in a big pile of figures, and finding the
average is often the first step towards seeing the bigger picture . With averages at
your disposal, you'll be able to quickly find the most representative values in your
data and draw important conclusions. In this chapter, we'lIlook at several ways to
calculate one of the most important statistics in town-mean, median, and mode
and you 'lI start to see how to effectively summarize data as concisely and usefully
as possible.
Welcom e to the H ealth Club
t\ com mon m casur c 01' average is the me a n
Mean ma rh
Dealing wirh unknowns
Back to the rn can
Back 10 the Hcalth Club
Everybod y wa s Kung Fu figh ting
G ur da ta has outli crs
Th e ou tliers did it
\Valercooler conversation
fi nd ing thc median
H ow to find the median in three step s:
Business is boornin g
The Little D ucklings swimming class
\Vhat wen t wrong with the rn ea n and med ian?
"\lh ill shou ld we do for dara like this?
The Mean Exposed
In trodu cing the mode
Th rce sreps Ior fincling the mod e
4·6
4·7
48
49
50
53
54
57
58
60
61
62
65
66
69
69
71
73
74-
xi
table of contents
meQSur1nb VClrlqbllltj Clnd spreqd
Power Ranges
Not everything's reliable, but how can you tell?Average s do a great job of giving you a typical value in your data set , but they don't
tell you the full story . OK, so you know where the center of your data is, but offen
the mean , median, and mode alone aren 't enough information to go on when you're
summarizing a data set. In this chapter , we'lI show you how to take your data skills
to the next level as we begin to analyze ranges and variation .
All three plcvet-s hcvethe scme cvercqe sccrefor Shoot lng, but r need scmewcry of choosing betwee n th em.Think yau can he lp>
oo
xii
'NaIlted: oll e playcr
\Ve need to corn pare player scores
Use th e ran ge to clifle renr iare between ciat a se ts
T he prob lern wit h ou tliers
"Ve neecl to ge t awa y from outlie rs
Q uartiles co me to th e reseue
T he int erquart ile ra nge excludes outl icrs
Q uartile anatorny
We 're no t j us t lim iteel to qu a rtiles
So wh at a re percenril es?
Box a nd wh isker plot s le t you visua lize ranges
Variab ility is m ore rhan just sp read
C a lcula ting ave rage disra nces
We ca n ca lc ula te va riar ion with the va ria neo . ..
. , .but sta nda rd deviat ion is a rnore intuitive measure
Standard De via tion Exposed
A quicker ca lc ularion for va ria nce
Whar if we need a ba seline for compari son?
Us e standard scores to compare va lues across data sers
Int erp reting sranda rd sco res
S ratsville All Stars win the leag ue!
84-
85
8G
89
9 1
92
93
94
98
99
JOO
104
105
lOG
107
108
11 3
118
11 9
120
125
table of contents
4calcl.l lett1n,g pl'9bablllt1es
Taking Chances
Life is full of uncertainty.Somet imes it can be impossible to say what will happen from one minute to the
next. But certain events are more likely to occur than others, and that 's where
probability theory comes into play. Probability lets you predict the future by
assessing how likely outcomes are, and knowing what could happen helps you
make informed decisions. In this chapter, you'lI find out more about probabil ity
and learn how to take control of the future !
135
156
157
128
129
132
16 1
167
169
170
172
173
181
182
183
159
136
142
147
148
149
155
EH Da n's Gr and Siam
Roll lIjJ for roule tte!
What a re the chances?
Find rou lette proba biliries
Y OLI can visualize probabiliries wirh a Venn diagram
Y OLI can also add prob abi lities
Exclusive cvents and inte rsecring evenrs
Proble ms at th e int ersect io n
So me more notanon
Another unl ucky spin . . .
Conclitions apply
Find conditioua l pro babilities
Tre cs also hclp YOll ca lculare cond itiona l probabi liries
H audy hint s for workin g with trees
Step I : Finding P(Black n Even)
Step 2: Fineling P(Even)
Step 3: Find ing P(Black 1Even)
Use the Law of Total Probability 1O find P(B)
lnrrodu cing Bayes' Theore m
If events a ffecr eac h other, the y are dependent
rf evcn ts do not affecr each orher, they are indcpendenr
More on calcu lating proba bility for independcnt evcnts
sN
0 0 0 119- i6
xiii
table ot contents
5uSlnh' dlscrete pr9bab'll!tj d'lstribut'i9n8
Manage Your Expectations
Unlikely events happen, but what are the consequences?So far we've looked at how probabilities tell you how likely certain events are. What
probability doesn't tell you is the overall impact of these events, and what it means
to you. Sure, you'lI sometimes make it big on the roulette table , but is it really worth it
with all the money you lose in the meantime? In this chapter, we'lI show you how you
can use probability to predict long-term outcomes, and also measure the certainty
of these predictions.
xiv
Back ar Fat Dan 's Casi no 198
We can compose a probab ility distr ibution for the slot machine 20I
Expect ation gives )'ou a pred iction of the results. . ; 20+
. . .and varia nre teils you about the sprcad o f the resul ts 205
Var iances and probabiliry distributions 206
Ler's calc ulare the slot machine 's varian ce 207
Fat Dan changed his prices 2 12
T herc 's a linear relationship between E(X) and E('r') 2 17
Siot machine transforrnation s 2 18
Gen eral forrn ulas lor linear tran sforms 219
Ever y pull of the lever is an ind ep cndent obscrvat iou 222
Obser vation sho rtcuts 223
New slot machine on rhe blec k 229
Adel E(X) and E(V) to ger E(X + V)... 230
... an d subtract E(X) and E(V) to get E(X - Y) 231
You can also add and subrract linear tran sformarions 232
] ackl)ot! 238
6
table of contents
permutat19ns and c9mblnatl9ns
Making Arrangements
Sometimes, order is important.Counting all the possible ways in which you can order things is time
consuming, but the trouble is, this sort of information is crucial for
calculating some probabilities. In this chapter, we'll show you a quick way
of deriving this sort of information without you having to figure out what all
of the possible outcomes are. Come with us and we'lI show you how to
count the possibilities.
The Statsville Derby 242
It's a three-liorse race 243
How many ways can they cross the finish line? 245
Calculate the nurnber of ilrrangements 246
Going round in cireles 247
It's time Ior the novelty race 251
AlTiInging by individuals is different than mranging by type 252
We need to mrange animals by type 253
Ceneralize a lormula for alTanging duplicates 25't
It's time for the tweury-horse race 257
How milny ways can we fill the top three positions? 258
Examining perrnutations 259
What if horse order doesn'r matre I' 260
Exarnining combinations 26 J
Cornbination Exposeel 262
Does order rcally matter? 262
It's the enel of the ra ce 268
xv
table of contents
7t:ß9metrlc. bln9mlctl . and p9'lss9n d'lstr'lbutl9ns
Keeping Things Discrete
Calculating probability distributions takes time.So far we've looked at how to calculate and use probability distributions, but wouldn 't it be
niee to have something easier to work with, or just quicker to calculate? In this ehapter,
we'lI show you some special probability distributions that follow very definite patterns .
Onee you know these patterns, you 'lI be able to use them to calculate probabilities,
expectations, and variances in record time. Read on, and we'lI introduee you to the
geometrie, binomial and Poisson distributions.
Popcorn machine
xvi
Drin'ks machine
We necd to lind Chads probab iliry d istr ibutio n
The re 's a pa trern ro rhi s p ro bability distr ibutio n
The p robabi lity di str iburion can be re p rese nred a lgc b rai ca lly
T hc ge o met rie d istr ibuti on a lso wo rks wit h ine quali iies
T he pattern of ex pccrat ions fo r the geo metr ie d istribu tion
Expecta tion is I1p
Find ing rhe var ia nce far our di st riburion
A quick gui de to th e ge o merric d istrib ution
Who Wilnt s to Win a Swivel Chair!
You 've masrered th e gcomerric distr ibution
Sho uld you play, 01' walk away?
Generillizin g the probabi liry fo r rhree questions
Ler's gc nera lize th e proba biliry further
Wha r's the expecrar io n a nd va riance?
Bino rn ial expecrat io n an d va riance
Your q uick gu ide to rhe biuornial d isrr ibu rio n
Expccrarion a nd va riance for the Po isson clisuibu rio n
So wha t's the proba biliry d istr ibutio n?
Combine Po isson va riables
The Poisson in d isg uise
Your q uick glli de ro the Po isso n distribution
273
274
277
279
280
281
283
284
287
287
291
293
296
298
301
302
308
312
313
316
31 9
8
table 01' contents
lIS1nh' the norma] dIstrIbutlon
Being Normal
Discrete probability distributions can't handle every situation.So far we've looked at probability distributions where we've been able to specify exact
values , butthis isn'tthe case for every set of data. Some types of data just don't fit the
probabil ity distr ibutions we've encountered so far. In this chapter, we'lItake a look at
how continuous probability distributions werk, and introduce you to one of the most
important probability distributions in town-the normal distribution.
°0
Discrere da ra ra kes exact values...
,. ,but not all nu meric da ta is discrete
Wh ar's the delay?
\\Ie ueed a probabiliry distrib urion for continuous data
Probability dcn sity Iuncti ons can be used for continuous da ta
Probability == area
Ta calcularc probabiliry, star t by finding ~x).
.. .rhcn find probability by fillCling the area
We've found rhe probabiliry
Sear clting for a soul mar e
Male modclling
T he normal distribution is <I n " idea l" model for continuous dara
So how da we find nor mal probab ilities?
Thrce steps to ca lcula ring normal probabilirics
Srep I: Derermine your distrib ution
Srep 2: Staudardizc to N (O , I )
To sta ndardize, first move the mean.
... then squash rhe wid th
Now find Z Ior rhe spccific value you want to find probability Ior
Ste p 3: Look up the probabil ity in your ha ndy tab le
326
327
328
329
330
331
332
333
337
338
339
340
341
34·2
343
344
345
345
346
349
/
"--'./~
.,.-- ......
,"
.r:":
/,,
xvii
table of contents
9uSlng the normal dlstrlbut19n II
Beyond Normal
If only all probability distributions were normal.Life can be so much simplerwith the normal distribution . Why spend all your time
working out individual probabilities when you can look up entire ranges in one swoop, and
stillleave time for game play? In this chapter, you 'll see how to solve more complex
problems in the blink of an eye, and you'lI also find out how to bring some of that normal
goodness to other probability distributions.
All aboard the Love Train 363
Normal br ide + normal g roo m 364·
!t's still j ust weight 365
How's th e comb ined weight dis trib uted? 367
Find ing probabil ities 370
M ore pe ople wa nt the Love Trai n 375
Lin ea r tran sforrn s descr ibe underl ying cha nges in va lues. . . 3 76
.. .a nd ind cpcndenr observarions dcscri bc how ma ny valu cs you have 377
Expc ctat ion and var iance for ind ependent ob servations 378
Should we play, 0 1' walk away? 383
Normal distribution to th e rcscu e 386
Wh en to approxirna re the b inom ial d istribu tion wi th the nor ma l 389
Rev isiring the normal approxi rna rion 394
L
_ - T he bino rnial is d iscrerc, but the normal is co nrinuous 395
Apply a conrinuity correc tion befor e cal cu la ting the a pprox imai ion 396
The No rmal Dist ribu tion Exposed 40 4
All ab oarr l the Love T rain 40 5
Wh en to approxirna te the binomia! clistributio u with rhe norma l 407
A runaway success! 4 13
xviii
x x+x~,}th ad.H. " J" ;"d<ft"d<"t."""rI~bo" ~ 'f.. .
~
x +x+x x+x+x+x
10
table of contents
us'lnh' stat'ltltkaJ samPllnh'
Taking Sampies
Statistics deal with data, but where does it come from?Some of the time, data's easy to collect, such as the ages of people attending a health
club er the sales figures for agames company. But what about the times when data isn't
so easy to coliect? Somet imes the number of things we want to coliect data about are so
huge that it's difficult to know where to start . In this chapter, we'li take a look at how you
can effectively gather data in the real world , in a way that's efficient , aceurate, and can
also save you time and money to boot. Welcome to the world of sampling.
T he Mighty G umball taste tcst
T hey're run ning out of gumballs
Test a gum billl sarnp le, no t rhe whoie gUlllball pop ularion
H ow sal11pling works
W hcn sarnpling goes wrong
H ow to desi gn a sa m ple
De fine your sampling frame
Sometimes sam ples ca n be bia sed
So urces of bias
H ow to choose your sa mple
Simple ra ndo m sarnpling
How to choose a simple random sarnple
T here are other types of sampling
We ca n use stra rified sam pling,..
... 0 1' \\'e ca n use clusrcr sarn pling.. .
.. .0 1' even systematic silmpling
t\'lighty G urn bal ! has a sample
4 16
417
4 18
419
420
422
423
424
425
4·30
430
431
432
432
433
+33
439
table of contents
xx
11est1mat1nh j9Ur p9pulQtJ9n
Making Predictions
Wouldn't it be great if you could tell what a population waslike, just by taking one sampie?Before you can claim full sam pie mastery, you need to know how to use your sampies
to best effect once you've collected them. This means using them to accurately predict
what the population will be like and coming up with a way of saying how reliable your
predictions are. In this chapter, we'll show you how knowing your sampie helps you
get to know your population, and vice versa .
So how long does flavor really last fo r? 442
Ler's sta rt by estirnating the popul ation mcan 44 3
Poin t estim a to rs ca n a pproximare po pul ation pa ra rnete rs 444
Ler's estinrare the popula tion variance 448
We nee d a different po in t estimaror tha n sa rnple va r ia nce 4'~9
W hieh formula 's whi eh? 451
It 's a quest ion of propo rt ion 454
So how do es this relate to sarnpling? 459
The sa m pling dist ribution o f prop ortion s 460
So wha r's the expe cta rion of Ps? 462
Ami whar's the va rianeo of P ? 46 3.'
Find th e d istribu tion of P 464
P follows a normal distribu tio n 465s
' Ne need pr obabilities fo r rhe sa rnple mean 4 71
The sarn pling distribut ion of the mean 472
Find the expccration for X 4 74
What about the th e va riance of X? 476
So how is X d ist ributed ? 480
If n is la rge, X can still be app roximared by the normal dist riburion 481
Using the cenrral limit theorem 482
12
table 0' contents
c9nstructlng c9nt'ldenoo 'lntery'qJs
Guessing with Confidence
Sometimes sampIes don't give quite the right result.You've seen how you can use point estimators to estimate the precise value of the
population mean, variance, or proportion, but the trouble is, how can you be certain that
your estimate is completely accurate? After all, your assumptions about the population
rely on just one sampie , and what if your sample's off? In this chapter, you'lI see another
way of estimating population statistics, one that allows for uncertainty. Pick up your
probability tables, and we'll show you the ins and outs of confidence intervals.
M ight y Gu ruball is in tro uble
The pro blern with precision
l ntro ducing con fidence intcrvals
Four steps Ior lindin g confiden ce intervals
Srep 1: Ch oose your population statistic
Srep 2: Find its sampling distribution
Srep 3: Decide on th e level of co nfidence
Sre p +: Find the confidence limits
Sta rt by finding Z
Rewrit c the incquality in terms of m
fi nally, find the value of X
Youve found the confidence inrerval
Lets surn ma rize the sreps
Handy shortcuts for confidence inte rvals
Step I: Choose your po pu larion sta tistic
Srep 2: fi nd its samp ling distr ibution
Step 3: D ecide on the level of confide nce
Stcp 4: Find the confidence lirnits
T he t-distrib urion vs. the no rma l distributio n
488
489
490
491
492
492
494
+96
497
498
50 1
502
503
504
508
509
512
513
515
xxi
table of contents
13uSlng hyp9theslStests
Look at the Evidence
Not everything you're told is absolutely certain.The trouble is, how do you know when what you're being told isn't right? Hypothesis
tests give you a way of using sampIes to test wheth er or not statistical claims are likely
to be true. They give you a way of weighing the evidence and testing whether extreme
results can be expla ined by mere co incidence, or whether there are darker forces at
work . Come with us on a ride through this chapter, and we'lI show you how you can use
hypothesis tests to confirm or allay your deepest suspicions.
xxii
Sra rsvilles new m iracle drug
Reso lving the conflict from 50,000 fec r
The six srcps for hypothesis testin g
St er I : D ecicle on the hyp oth csis
S rep 2: Choo se your test sta tistic
St er 3: D ere r m ine thc cri tica l region
St ep 4: Find the p-value
Srep 5: Is the sa rnp le result in the critical reg ion?
Step 6: M ake your c1 ec.isio n
W hat if the sa rnple size is la rger?
Le t's co ncluct a nother hypo th esis test
Stcp I : Decide on the hypothescs
St er) 2: Choose the rest sta tistic
U se the normal to ap proximate the binomial in our test sta tisric
St e r) 3: F ind rhe crit ica l region
Let's sta rr with T yp e I errors
What abour Type II er ro rs?
F incling e rr ors for Sn oreC ull
'Ne need ro lincl th e ra nge of val ues
Fincl P(T yp e II e rror)
Introclucing power
5'22
526
5'27
528
531
532
535
537
537
540
543
543
544
547
548
556
557
558
559
560
56l
14
table of contents
the XZ dlstrlbut19n
There's Something Going On...
Sometimes things don't turn out quite the way you expect.When you model a situat ion using a particu.ar probability distr ibution , you have a
good idea of how things are likely to turn out lonq-terrn . But what happens if there are
differences between what you expect and what you get? How can you tell whether
your discrepancies come down to normal f1uctuations, or whether they're a sign of
an underlying problem with your probab ility model instead? In this chapter , we'll
show you how you can use the X2 distribution to analyze your results and sniff out
suspicious results.
T here may be trouble a head ar Fat Dan 's Cas ino
Let's srart wirh rhe slot rnachines
The '1/ tcst asscsses diflcrence
So whar cloes rhc test starisric represent?
Two ma in uses o f rhe '1/ d isrr iburion
V represen ts elegrees of Irccdorn
Whar's rhe significance?
H ypoth esis tesring with X~
You 've solveel the slot rnachine mystery
Fat Dan has anorher problern
T he X2 d isrributio n ca n tcst Ior ind epe ndence
You can finel rhe expccted Irequencies using probab ility
So what are rhe (rcquc ucics?
We still nced to calculare degrces of Ircedom
Gene ralizing the c1 egree s of Ireedom
A IlCI the lormula is...
You've savcd the casino
568
569
571
572
573
574
575
576
579
585
586
587
~)88
59l
596
597
599
xxiii
table of contents
15correlarlcn emd regresS19n
What's My Line?
Have you ever wondered how two things are connected?So far we've looked at stausfies that tell you about just one variable-like rnen's height,
points scored by basketball players, or how long gumball flavor lasts-but there are othet
statistics that tell you about the connection between variables. Seeing how things are
connected can give you a lot of information about the real world, informat ion that you can
use to your advantage . Stay with us while we show you the key to spotting connectio
correlation and regression .
Let's analyze sunshine and atrendanc e 60 7
Exploring tw es of data 608
Visualizing bivar iare da ra 609
Scatter d iagr ams show you patre rns 6 12
C orrelarion vs. ca usano n 614
Predict values wirh a line of besr fit 618
Your bes t gue ss is still a guess 6 19
We need to rninirnize the errors 620
Introdu cing the surn of sq uared err ors 62 1
Find rhe equation for the line of best fit 622
Fincling the slope for the line of best fit 623
Fincling the slope for rhe line of besr fit, coniinucd 624
Wc've fou nd b, bur whar abo ut a? 625
You've m ad e the connection 62 9
Let's look ar some co rrela rions 630
T he correlation coel1icien t measures how weil the line fits rhe data 63 1
T here's a formu la for calcularing the corre lation coeflicicnr, r 632
Find r for the concert dara 633
Find r for the co neert daia , con tinued 634
xxiv
Fee! that funkyrh ythm . baby.
•1
••11
table ot contents
lett9v'ers
The Top Ten Things (we didn't cover)
Even after all that, there's a bit more. The re are just a few more
things we think you need to know. We wouldn 't feel right about ignoring them ,
even though they only need abrief mention . So before you put the book down,
take a read through these short but important statistics tidbits .
# I. O rher ways of prese n ring dara 644
# 2. Distr ibution ana to rny 645
# 3. Experiments 646
#4. Least sq ua re reg ression alterna te notation 648
#5 . The coeffi cient of determin arion 649
# 6. Non-linear relati on ships 650
#7 . The co nfide nce inrerva l Ior the slope of a regression line 65 1
#8 . Sa mpling disrributions - the differen ce between two rnean s 652
# 9. Sa mpling distriburions - the d iflerence be tween two proport ion s 653
# 10 . E(X) and Var(X) for continuous probability d istribu tions 654
stattstics telbles
Looking Things up
Where would you be without your trusty probability tables?Understanding your probabil ity distributions isn't quite enough . For some of them , you
need to be able to look up your probabilities in stand ard probabil ity tables. In this
appendix you 'll find tables for the normal , t and X2 distributions so you can look up
probabilities to your heart's content.
St andard normal prob abi lities
t-distriburion cr itical value s
x' cri tica l values
658
660
66 l
xxv